Hsrproceedings.org

Caring for the Critically Ill Patient
Effect of Prolonged MethylprednisoloneTherapy in Unresolving AcuteRespiratory Distress Syndrome G. Umberto Meduri, MD; A. Stacey Headley, MD; Emmel Golden, MD; Stephanie J. Carson, RN; Reba A. Umberger, RN; Tiffany Kelso, PharmD; Elizabeth A. Tolley, PhD Context.—No pharmacological therapeutic protocol has been found effective in
modifying the clinical course of acute respiratory distress syndrome (ARDS) and Objective.—To determine the effects of prolonged methylprednisolone therapy
on lung function and mortality in patients with unresolving ARDS.
injury to the terminal respiratory unitswith exudative pulmonary edema.1 Al- Design.—Randomized, double-blind, placebo-controlled trial.
Setting.—Medical intensive care units of 4 medical centers.
vive the initial insult that precipitates Participants.—Twenty-four patients with severe ARDS who had failed to
improve lung injury score (LIS) by the seventh day of respiratory failure.
Interventions.—Sixteen patients received methylprednisolone and 8 received
placebo. Methylprednisolone dose was initially 2 mg/kg per day and the duration of treatment was 32 days. Four patients whose LIS failed to improve by at least 1 point after 10 days of treatment were blindly crossed over to the alternative treat- Main Outcome Measures.—Primary outcome measures were improvement in
tense inflammatory and fibrotic activitywith maladaptive lung repair.1 Mortal- lung function and mortality. Secondary outcome measures were improvement in multiple organ dysfunction syndrome (MODS) and development of nosocomial in- Results.—Physiological characteristics at the onset of ARDS were similar in both
groups. At study entry (day 9 [SD, 3] of ARDS), the 2 groups had similar LIS, ratios For editorial comment
of PaO2 to fraction of inspired oxygen (FIO2), and MODS scores. Changes observed see pp 181 and 182.
by study day 10 for methylprednisolone vs placebo were as follows: reduced LIS(mean [SEM], 1.7 [0.1] vs 3.0 [0.2]; PϽ.001); improved ratio of PaO2 to FIO2 (mean[SEM], 262 [19] vs 148 [35]; PϽ.001); decreased MODS score (mean [SEM], 0.7 ses7,8 of randomized trials investigating a [0.2] vs 1.8 [0.3]; PϽ.001); and successful extubation (7 vs 0; P = .05). For the short course (Յ48 hours) of high-dose treatment group vs the placebo group, mortality associated with the intensive care unit was 0 (0%) of 16 vs 5 (62%) of 8 (P = .002) and hospital-associated mortality was 2 (12%) of 16 vs 5 (62%) of 8 (P = .03). The rate of infections per day of treat- effect. In contrast, we and others have re- ment was similar in both groups, and pneumonia was frequently detected in the function during prolonged methylpred- Conclusions.—In this study, prolonged administration of methylprednisolone in
nisolone administration in medical9-13 and patients with unresolving ARDS was associated with improvement in lung injury and related with improvement in lung func-tion. In phase 2 trials involving 34 pa-tients, we reported mortalities of 17% in From the University of Tennessee Lung Research College of Medicine, and the College of Pharmacy Program, the Baptist Memorial Hospitals (Drs (Dr Kelso), University of Tennessee, Memphis.
Meduri, Headley, and Golden and Mss Carson and Reprints: G. Umberto Meduri, MD, Division of Pulmo- Caring for the Critically Ill Patient section editor: Umberger), the Veterans Affairs Medical Center (Dr nary and Critical Care Medicine, University of Tennes- Deborah J. Cook, MD, Consulting Editor, JAMA. Advi- Meduri), and the Departments of Medicine (Drs see College of Medicine, 956 Court Ave, Room H314, sory Board: David Bihari, MD; Christian Brun-Buisson, Meduri, Headley, and Golden and Mss Carson and Memphis, TN 38163 (e-mail: [email protected] MD; Timothy Evans, MD; John Heffner, MD; Norman Umberger) and Preventive Medicine (Dr Tolley), Methylprednisolone Therapy in ARDS—Meduri et al 1998 American Medical Association. All rights reserved.
if a patient crossed over into the alter- nate treatment arm or exited the study.
nisolone treatment, lung histology showed to limit plateau pressure at 35 cm or less of water.18 Unless contraindicated,19 di- monitored daily for the development of in- agnostic fiberoptic bronchoscopy with bi- fections or other complications. If a pa- sponders.11 These findings suggested that the efficacy of prolonged methylpredniso- ventilation to exclude occult ventilator- increase in minute ventilation of more than ably had greater than a 1-point reduction nostic criteria followed consensus guide- lines.19 Febrile patients received a previ- related infection, urinary tract infection, quired appropriate antibiotic therapy for ylprednisolone therapy could be effective 3 or more days prior to study entry.
sis of infection(s) was established by strict Data Collection and
Outcome Definitions
domized, double-blind, placebo-controlled classified as either direct or indirect lung trial of patients with unresolving ARDS.
injury. Direct lung injury was defined as The primary objective was to test the hy- Treatment Protocol
lowing data were obtained on days 1, 2, 3, 5, and 7 of ARDS and on days 1, 3, 5, 7, 10, day from day 1 to day 14, 1 mg/kg per day of the LIS (while intubated),23 MODSscore,24 systemic inflammatory response day from day 22 to day 28, 0.25 mg/kg perday on days 29 and 30, and 0.125 mg/kg Patient Selection, Management,
per day on days 31 and 32. If the patient and Randomization
1 point.11 Resolution of individual organ tablets, which were similar in shape, size, and color to methylprednisolone tablets.
vision for patients who did not respond to adjudication of death with refractory hy- gible if they met all of the following cri- at least 1 point after 10 days of treatment were blindly crossed over to the alterna- ventilation with an LIS of 2.5 or greater posure to ineffective therapy; this alter- Statistical Analysis
for 32 days. Patients exited the study if tial clinical trial with nonconstant inspec- ing requiring transfusion, had Candida deaths observed (ie, after 3 and 5 deaths species recovered from multiple sites, or sive burns, life expectancy of less than 3 months because of terminal illness, preg- Methylprednisolone Therapy in ARDS—Meduri et al 1998 American Medical Association. All rights reserved.
was adopted instead of the Whitehead tri- exact tests. Total sample sizes, 1-tailed P with standard care alone. In a sequential ing accumulated data) for the first inspec- Pregnancy: 1Disease Process Requiring Methylprednisolone: 1 clinical trial, the goal is to stop the trial tion were 14, .03, and 0.63, respectively, Participation in Another Investigational Study: 1 and for the second inspection were 24, .002, mortality. In a 1-sided scenario, the trial lar test under the presumption of 1 hypo- tial triangular test, the decision to reject the null hypothesis had been made at a sig- nificance level of less than .05 and a power level of greater than 0.95. Subsequently, despite a total sample size of only 24 pa- were no protocol violations; all data were Figure 1.—Flow diagram showing the progression tailed or 2-tailed tests were used, or ␾ mission to the trial was determined by se- analyzed each clinical variable used to cal- quential analyses of results as data were accumulated.16 In this sequential clinical 95% confidence intervals were estimated.
trial, the sample size was not a fixed num- nonsignificant difference only for a lower ber. The actual number of subjects was de- [30] vs 290 [49]; P = .09). In the methyl- test. For survival curves, censored times lar test of Whitehead16 was used for test- methylprednisolone and placebo groups.
ing differences between the 2 groups. The (P = .19). In the methylprednisolone group, working levels of ␣ and ␤ were .05. Thus, were compared by the Student t test and the working level of power was 0.95. Dur- the Mann-Whitney U test. For continu- ing the design of this trial, we postulated cebo group (P = .13). Twelve patients had contrasts in the context of repeated-mea- nisolone group would both be 0.80 and that sures analysis of variance (split plot).26 ration (placebo group). Ten infections were would both be 0.50. A fixed sample size of to control for potential confounding vari- 99 patients was required to detect a dif- ables (ie, all 30 variables used to calculate (Table 1), and all patients received anti- ference of 0.30 between the proportion of biotic treatment for at least 3 days prior 2-sided alternative hypotheses) using Co- 10 days of the study is shown in Figure 2 treated with standard care (ie, placebo).
formed with strata defined by severity of boundaries were constructed to satisfy the ratio of PaO2 to FIO2 by day 5 (161 [13] to working levels of ␣ and ␤. The decision to 217 [16]; P = .01), static lung compliance end the trial was made when the test sta- by day 7 (25 [1] to 32 [2]; P = .002), LIS by day 5 (3.0 [.01] to 2.2 [.01]; PϽ.001), mean [1] to 22 [3]; P = .04), and MODS score24 by day 7 (1.7 [.01] to 1.2 [.01]; PϽ.001). None of these variables improved in the placebo cebo. Data were reported as mean (SE).
group. During the first 10 days of treat- Both groups had similar clinical and physi- ment, the percentage of circulating imma- ological23-25,27 characteristics at onset of tion did not change in either group. In the Methylprednisolone Therapy in ARDS—Meduri et al 1998 American Medical Association. All rights reserved.
Table 1.—Clinical and Physiological Characteristics at Onset of ARDS* Characteristics
Methylprednisolone
At Onset of ARDS
At Study Entry
*ARDS indicates acute respiratory distress syndrome; NA, not applicable; APACHE, Acute Physiology and Chronic Health Evaluation; FIO2, fraction of inspired oxygen; MODS, multiple organ dysfunction syndrome; PEEP, positiveend-expiratory pressure; and IQR, interquartile range.
Figure 2.—Mean (SE) changes in lung injury score †Indicates APACHE III score on admission to intensive care unit.27 ‡Causes of direct injury to the lung include 11 bacterial pneumonias (8 community acquired, 4 in each group), 3 chemical aspirations (2 in patients randomized to placebo), and 1 pulmonary blastomycosis (patient randomized to (FIO2), and multiple organ dysfunction syndrome placebo). Causes of indirect injury to the lung include 5 extrapulmonary sepsis, 2 postoperative ARDS, and 2 drug (MODS) score during the first 10 days of treatment reactions (1 tricyclic antidepressant overdose and 1 anaphylactic reaction to urokinase [patient randomized to in the methylprednisolone group and placebo group. Error bars indicate SEs. There was no sta- §Nosocomial infections diagnosed by strict criteria within 5 days of randomization in the methylprednisolone group tistical difference between the methylprednisolone include bacteremia, fungemia, sinusitis, wound infection, 2 urinary tract infections, and 2 ventilator-associated and placebo groups at the time of entry into the pneumonias; the placebo group had 2 catheter-related infections.
study. The values on day 1 were obtained priorto initiating treatment. In the methylprednisolonegroup, a statistically significant improvement was but 1 within 4 days of removal of mechani- achieved for the ratio of PaO2 to FIO2 on day 5 cal ventilation. One patient in the meth- (PϽ.01), LIS on day 5 (PϽ.001), and MODS score served by day 3 (P = .004) and an increase on day 7 (PϽ.001). In the placebo group, no statis- in total leukocyte count by day 7 (P = .04).
tically significant improvement was achieved duringthe first 10 days of treatment. The number of patients in the methylprednisolone groups on study days 7 and 10 were 14 and 9, and in the placebo group, 6 and 6, respectively. The asterisk indicates were crossed over to the placebo (as dic- were correlated (r = 0.688 [0.165]).
and Table 3. After controlling for poten- trial ended early, a sufficient number of tial confounding variables, differences in mained significant (all P values Ͻ.008).
observed in all patients initially random- lone after 10 days of placebo (P = .04).
because of candidemia with positive find- lation is shown in Table 3. For survivors tion rate ratio per day of mechanical ven- photericin treatment and was extubated.
Methylprednisolone Therapy in ARDS—Meduri et al 1998 American Medical Association. All rights reserved.
Table 2.—Outcome Measures on Study Day 10* Outcome Measures
Methylprednisolone
Patients with Ͼ1-point reduction in LIS, No. (%) Crossed over because of failure to improve LIS† Infections per 100 patient-days of treatment࿣ *Data are reported as absolute or mean (SEM). NA indicates not applicable; FIO2, fraction of inspired oxygen; and MODS, multiple organ dysfunction syndrome.
†Four patients randomized to placebo failed to reduce lung injury score of 1 point or more from study entry and were blindly crossed over to methylprednisolone. Two patients randomized to placebo died before study day 10.
‡Pulmonary artery pressure values are reported for study day 7.
Survival Curve for Methylprednisolone Group §Improvement was significant for platelet count by day 5 (P = .004), serum creatinine by day 7 (P = .04), and serum ࿣Number of infections divided by number of treatment days received and multiplied by 100.
Censored Times for Survivors in Methylprednisolone Group (n=14) Outcome Measures
Methylprednisolone
Death associated with unresolving ARDS, No.† MODS-free days by study day 28, mean (SEM)‡ Figure 3.—Survival curves of patients receiving Duration of mechanical ventilation, median, d methylprednisolone and placebo. In the methyl-prednisolone and placebo groups, survival times for *ICU indicates intensive care unit; ARDS, acute respiratory distress syndrome; NA, not applicable; and MODS, multiple organ dysfunction syndrome.
14 and 3 patients, respectively, were classified as †ARDS failed to resolve and the patient required more than 0.8 fraction of inspired oxygen (FIO2) to maintain a PaO2 of more than 60 mm Hg at the time of death.
‡Resolution of individual organ dysfunction followed expert panel recommendations23 and included cardiovascular system, a systolic blood pressure greater than 90 mm Hg; respiratory system, a ratio of PaO2 to FIO2 greater than and failure to improve (P = .04). In agree- 400; nervous system, a Glasgow coma score of 15 or greater; coagulation, a platelet count greater than 120 × 109/L(120 000 µL); renal system, a serum creatinine level less than 133 µmol/L (1.5 mg/dL); and hepatic system, a bilirubin level less than 21 µmol/L (1.2 mg/dL).24 viously provided data to support a causalrelationship between intensity and du- large bleeding rectal ulcer requiring ma- lone therapy should be started before fi- produced in the lung into the systemic cir- acellular fibrosis in which type I colla- effect a containment of the host defense re- sponse, which is crucial to the reversal of spond initially to the alternate treatment and procedures for infection surveillance, ported by previous animal32-34 and clini- cal9,13,14,35,36 studies. In experimental acute BAL. Using a sequential clinical trial, we tion has been effective in decreasing lung albumin; normalization of gallium citrate score and with a significant reduction in influence.33,34,37 Furthermore, a short course histology); and a lower mortality rate.11,12 to the first 6 days of experimental acute was associated with a 50% failure rate; a polysaccharide challenge in humans is sig- nificantly enhanced by a prior short course Methylprednisolone Therapy in ARDS—Meduri et al 1998 American Medical Association. All rights reserved.
Table 4.—Complications Observed During Therapy* extent that the treatment effect mightmerely have reflected some confounding Complications
Methylprednisolone*
variable, such as severity of illness.
the placebo group and did die. On the day *Data are expressed as No. (%). There was no statistically significant difference in any variable among the 2 groups except for reduction in hemoglobin of more than 0.20 g/L vs study admission (P = .03).
†In the methylprednisolone group, 4 of these infections were identified after intensive care unit discharge. In the placebo group, 4 of these infections (2 Staphylococcus epidermidis bacteremias, 1 Klebsiella empyema, and 1 Clostridium difficile colitis) developed after crossover to methylprednisolone.
ences in outcome between the 2 groups.
‡Etiology of ventilator-associated pneumonias in the methylprednisolone group included 3 Staphylococcus aureus, 2 Pseudomonas aeruginosa, 1 Acinetobacter, 1 Klebsiella pneumoniae, 1 Escherichia coli, and 1 Entero- bacter; and in the placebo group, 1 Pseudomonas aeruginosa.
§Etiology of sinusitis included 1 Staphylococcus aureus and 1 Staphylococcus aureus and Proteus mirabilis.
࿣Other infections in the methylprednisolone group included 1 lung abscess and 1 infected intra-abdominal tions in 7 prior observational studies9-15 hematoma (both requiring surgical drainage); the placebo group included 1 empyema and 1 Clostridium difficile ARDS. Our findings suggest that, in pre-vious large, randomized, multicenter of glucocorticoids,38 and this response may explain the difference in infection-related course, 24-hour treatment39 and the find- frequently develops in the absence of fe- ver, and infection surveillance, including weekly surveillance bilateral BAL, is re- adrenal axis, glucocorticoid receptor func- and other serious infections.11 In the cur- tion, and cytokine modulation of the host rent study, the infection rate was not sig- defense response in critical illness sug- mechanical ventilation. None of the noso- comial infections with potentially lethal We are indebted to Harold Dickson, PhD, and Sandra Tsiu, PharmD, of the Baptist Health Care Foundation, Memphis, Tenn, for their invaluable help in initiating this study; to the following physi- cians for allowing us to study their patients: Johnny reinforces the findings of Headley et al6 Belenchia, MD, Albert Chinn, MD, Lisa Kennedy,MD, Roy Fox, MD, Kenneth V. Leeper, MD, Muham- that appropriately treated nosocomial in- mad Zaman, MD, and Glen Williams, MD; to the Fel- fections, albeit a frequent complication of lows of the Pulmonary and Critical Care Division, ARDS, do not themselves cause death.
Department of Medicine, University of Tennessee, Memphis, for their generous help; to Julia Campbellfor assistance in data management; to the APACHE Medical System Inc, McLean, Va, for assistance in any trial early biases the estimate of the the calculation of the APACHE III score; and to John treatment effect. In this trial estimates Griffin, MD, Dennis Schaberg, MD, and David of the positive effects of methylprednis- Armbruster, PhD, for helpful critique of this article.
Special thanks go to our patients Janet Machala and This study was supported by the Baptist Memo- transcriptional activation of several cy- rial Health Care Foundation, Memphis, and by the University of Tennessee, Clinical Research Center,Memphis.
References
1. Meduri GU. The role of the host defense response
in the progression and outcome of ARDS: patho- physiological correlations and response to glucocor- ticoid treatment. Eur Respir J. 1996;9:2650-2670.
2. Kollef MH, Schuster DP. The acute respiratory
distress syndrome. N Engl J Med. 1995;332:27-37.
tor, and nitric oxide (3 key substances in 3. Krafft P, Fridrich P, Pernerstorfer T, et al. The
Methylprednisolone Therapy in ARDS—Meduri et al 1998 American Medical Association. All rights reserved.
acute respiratory distress syndrome: definitions, se- and clinical trial coordination. Am J Respir Crit lated to the severity of acute injury. Am Rev Respir verity and clinical outcome: an analysis of 101 clini- cal investigations. Intensive Care Med. 1996;22: 18. Slutsky AS. Mechanical ventilation.Chest. 1993;
32. Jones RL, King EG. The effects of methylpred-
nisolone on oxygenation in experimental hypoxemic 4. Bernard GR, Luce JM, Sprung CL, et al. High-
19. Meduri GU, Chastre J. The standardization of
respiratory failure. J Trauma. 1975;15:297-303.
dose corticosteroids in patients with the adult res- bronchoscopic techniques for ventilator-associated 33. Hesterberg TW, Last JA. Ozone-induced acute
piratory distress syndrome. N Engl J Med. 1987; pneumonia. Chest. 1992;102(suppl):557S-564S.
pulmonary fibrosis in rats—prevention of increased 20. Reddy R, Meduri GU, Stanley T, Umberger R,
rates of collagen synthesis by methylprednisolone.
5. Bone RC, Maunder R, Slotman G, et al. An early
Headley S. Bilateral BAL to diagnose ventilator- Am Rev Respir Dis. 1981;123:47-52.
test of survival in patients with the adult respira- associated pneumonia (VAP) in ARDS [abstract].
34. Hakkinen PJ, Schmoyer RL, Witschi HP. Po-
tory distress syndrome: the PaO2/FIO2 ratio and its Am J Respir Crit Care Med. 1997;155:A393.
tentiation of butylated-hydroxytoluene-induced response to conventional therapy. Chest. 1989;96: 21. Meduri GU, Mauldin GL, Wunderink RG, et al.
acute lung damage by oxygen: effects of predniso- Causes of fever and pulmonary densities in patients lone and indomethacin. Am Rev Respir Dis. 1983; 6. Headley AS, Tolley E, Meduri GU. Infections and
with clinical manifestations of ventilator-associated the inflammatory response in acute respiratory dis- pneumonia. Chest. 1994;106:221-235.
35. Briegel J, Kellermann W, Forst H, et al. Low-
tress syndrome. Chest. 1997;111:1306-1321.
22. Tornatore KM, Logue G, Venuto RC, et al. Phar-
dose hydrocortisone infusion attenuates the sys- 7. Lefering R, Neugebauer EAM. Steroid contro-
macokinetics of methylprednisolone in elderly and temic inflammatory response syndrome. Clin In- versy in sepsis and septic shock: a meta-analysis.
young healthy males. J Am Geriatr Soc. 1994;42: Crit Care Med. 1995;23:1294-1303.
36. Bollaert PE, Charpentier C, Levy B, et al. Re-
8. Cronin L, Cook DJ, Carlet J, et al. Corticosteroid
23. Murray JF, Matthay MA, Luce JM, Flick MR.
versal of late septic shock with supraphysiological treatment for sepsis: a critical appraisal and meta- An expanded definition of the adult respiratory doses of hydrocortisone. Crit Care Med. 1998;26: analysis of the literature. Crit Care Med. 1995;23: distress syndrome. Am Rev Respir Dis. 1988;138: 37. Kehrer JP, Klein-Szanto AJP, Sorensen EMB,
9. Hooper RG, Kearl RA. Established ARDS
24. Bernard GR, Doig G, Hudson LD, et al. Quan-
Pearlman R, Rosner MH. Enhanced acute lung dam- treated with a sustained course of adrenocorticol tification of organ failure for clinical trials [abstract].
age following corticosteroid treatment. Am Rev steroids. Chest. 1990;97:138-143.
Am J Respir Crit Care Med. 1995;151:A323.
Respir Dis. 1984;130:256-261.
10. Meduri GU, Belenchia JM, Estes RJ, Wunder-
25. Bone RC, Balk RA, Cerra FB, et al. Definitions
38. Barber AE, Coyle SM, Marano MA, et al. Glu-
ink RG, El Torky M, Leeper KV Jr. Fibroprolifera- for sepsis and organ failure and guidelines for the cocorticoid therapy alters hormonal and cytokine tive phase of ARDS: clinical findings and effects of use of innovative therapies in sepsis. Chest. 1992; responses to endotoxin in man. J Immunol. 1993; corticosteroids. Chest. 1991;100:943-952.
11. Meduri GU, Chinn AJ, Leeper KV, et al. Corti-
26. Sokal RR, Rohlf FJ. Comparisons among means:
39. Bone RC, Fisher CJ Jr, Clemmer TP, Slotman
costeroid rescue treatment of progressive fibropro- planned comparisons. In: Biometry. 2nd ed. New GJ, Metz CA, and the Methylprednisolone Severe liferation in late ARDS: patterns of response and York, NY: WH Freeman & Co; 1981:232-242.
Sepsis Study Group. Early methylprednisolone predictors of outcome. Chest. 1994;105:1516-1527.
27. Knaus WA, Wagner DP, Draper EA, et al. The
treatment for septic syndrome and the adult respi- 12. Meduri GU, Headley S, Tolley E, Shelby M,
APACHE III prognostic system: risk prediction ratory distress syndrome. Chest. 1987;92:1032-1036.
Stentz F, Postlethwaite A. Plasma and BAL cyto- of hospital mortality for critically ill hospitalized 40. Bamberger CM, Shulte HM, Chrousos GP. Mo-
kine response to corticosteroid rescue treatment in adults. Chest. 1991;100:1619-1636.
lecular determinants of glucocorticoid receptor late ARDS. Chest. 1995;108:1315-1325.
28. Meduri GU, Headley S, Kohler G, et al. Persis-
function and tissue sensitivity to glucocorticoids.
13. Hooper RG, Kearl RA. Established adult res-
tent elevation of inflammatory cytokines predicts a piratory distress syndrome successfully treated poor outcome in ARDS: plasma IL-1␤ and IL-6 are 41. Marx J. How glucocorticoids suppress immu-
with corticosteroids. South Med J. 1996;89:359-364.
consistent and efficient predictors of outcome over nity. Science. 1995;270:232-233.
14. Ashbaugh DG, Maier RV. Idiopathic pulmonary
time. Chest. 1995;107:1062-1073.
42. Barnes PJ, Adcock IM. Glucocorticoid recep-
fibrosis in adult respiratory distress syndrome: diag- 29. Meduri GU, Kohler G, Headley S, Tolley E,
tors. In: Crystal RG, West JB, Weibel ER, et al, eds.
nosis and treatment. Arch Surg. 1985;120:530-535.
Stentz F, Postlethwaite A. Inflammatory cytokines The Lung: Scientific Foundations. 2nd ed. Philadel- 15. Biffl WL, Moore FA, Moore EE, Haenel JB,
in the BAL of patients with ARDS: persistent el- phia, Pa: Lippincott-Raven Publishers; 1997:37-56.
McIntyre RC, Burch JM. Are corticosteroids sal- evation over time predicts poor outcome. Chest. 43. Kew KH, Ludwig EA, Milad MA, et al. Gender-
vage therapy for refractory acute respiratory dis- based effects on methylprednisolone pharmacoki- tress syndrome? Am J Surg. 1995;170:591-596.
30. Meduri GU, El Torky M, Winer-Muram HT. The
netics and pharmacodynamics. Clin Pharmacol 16. Whitehead J. The Design and Analysis of Se-
fibroproliferative phase of late adult respiratory quential Clinical Trials. New York, NY: John distress syndrome. Semin Respir Infect. 1995;10: 44. Meduri GU, Headley EA, Tolley A, et al. Plasma
and BAL procollagen type I and III levels during 17. Bernard GR, Artigas A, Brigham KL, et al. The
31. Shen AS, Haslett DC, Feldsien C, Henson PM,
ARDS and in response to prolonged methylpred- American-European Consensus Conference on Cherniack RM. The intensity of chronic lung inflam- nisolone treatment [abstract]. Am J Respir Crit ARDS: definitions, mechanisms, relevant outcomes, mation and fibrosis after bleomycin is directly re- Methylprednisolone Therapy in ARDS—Meduri et al 1998 American Medical Association. All rights reserved.

Source: http://www.hsrproceedings.org/consensus/allegati_topicsel/pdf_improve_survival/1998_JAMA_Meduri_Methylprednisolone.pdf